New media promotion optimization system and method based on Internet

Abstract

The invention discloses a new media promotion optimization system and method based on the Internet, and relates to the technical field of information promotion. The system comprises a multi-dimensional data analysis module, an intelligent content creation module, a user positioning promotion module, a user definition secondary promotion module and an effect evaluation adjustment module. The technical key points are as follows: the system introduces a rule engine to define active users and inactive users, so that the system can be flexibly adjusted according to different service requirements and user behavior changes; according to the method, the effect of each promotion activity can be comprehensively and accurately measured by setting a clear evaluation index and an evaluation formula, the most representative index is selected for evaluation according to different promotion modes, and the accuracy of an evaluation result is ensured; according to the comparison between the effect evaluation index and the preset standard threshold value, the system can automatically adjust the promotion frequency in real time, so that the waste of resources is avoided, the continuous optimization of the promotion activity is ensured, and the promotion efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of information promotion, and in particular to a new media promotion optimization system and method based on the Internet. Background Art

[0002] Information promotion is the process of conveying specific information to the target audience through various channels and means; this process aims to expand the coverage of information, enhance the influence of information, and encourage the target audience to be interested in, recognize, understand or take certain actions on the information; effective information promotion requires comprehensive consideration of multiple factors, including the characteristics of the target audience, the nature of the information, the choice of promotion channels and the formulation of promotion strategies, etc. At the same time, information promotion also needs to continue to innovate and adapt to market changes to ensure the effectiveness and sustainability of promotion activities; in actual applications, information promotion is widely used in various fields, such as corporate marketing, brand promotion, product promotion, public welfare promotion, etc. Through effective information promotion, enterprises or individuals can better convey information, expand influence, and achieve corresponding goals.

[0003] However, in traditional new media promotion, usually only the Internet technology is used to collect user information, and then the user portrait is obtained after analysis, and targeted promotion is carried out according to the user portrait. In this process, although the promotion is targeted, the promotion effects for different types of users are different. It is only based on the user portrait for conventional information promotion or release, and it is impossible to deal with different types of users in a more detailed manner. For example, users with different levels of activity, the effect achieved by a single and conventional promotion method is limited. While the promotion efficiency cannot be improved, it also wastes promotion resources to a certain extent. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In view of the deficiencies of the prior art, the present invention designs a new media promotion optimization system based on the Internet, and solves the problems raised in the background technology by running the system.

[0006] (II) Technical solution

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0008] A new media promotion optimization system based on the Internet, comprising:

[0009] Multi-dimensional data analysis module, which collects and analyzes multi-source data on the Internet to build user portraits;

[0010] Intelligent content creation module, based on user portraits, uses NLP and GAN technology to generate corresponding promotional content;

[0011] User positioning promotion module, based on user portraits and historical promotion data, introduces a deep learning model to predict the user groups interested in the promotion content;

[0012] The user-defined secondary promotion module builds a rule engine based on key indicators that reflect user activity, identifies the corresponding users as active or inactive users, and uses the method set in the solution database to promote according to the user type;

[0013] The effect evaluation and adjustment module evaluates the effect of each promotion under the promotion method corresponding to the user type, and adjusts the established promotion frequency in real time based on the generated effect evaluation index.

[0014] Furthermore, the constructed user portrait includes but is not limited to user interest preferences, browsing habits and interaction history.

[0015] Furthermore, key indicators include login frequency, usage time, and interactive behavior.

[0016] Furthermore, the process of building a rule engine to identify the corresponding user as an active user or an inactive user is as follows:

[0017] Set the AND logic mechanism in the rule engine and set the conditions as follows:

[0018] Daily active users: users log in at least once a day;

[0019] Weekly active users: users log in at least 3 times a week, and each session lasts more than 20 minutes;

[0020] Monthly active users: users make at least 5 purchases per month;

[0021] When the corresponding user meets all conditions within a given time period, he is marked as an active user; otherwise, he is marked as an inactive user.

[0022] Furthermore, when the promotion is performed by calling the method set in the solution database according to the user type, the solution database includes interactive content creation promotion and AI user behavior prediction promotion;

[0023] When the user type is: active user, the promotion method used is interactive content creation promotion;

[0024] When the user type is: inactive user, the promotion method used is AI user behavior prediction promotion.

[0025] Furthermore, the evaluation process of each promotion effect is as follows:

[0026] When interactive content creation and promotion is adopted, the first evaluation index is collected, including click-through rate, conversion rate and user satisfaction rate, and an evaluation formula is established based on the first evaluation index to calculate the effect evaluation index Tg;

[0027] The evaluation formula established based on the first evaluation indicator is as follows:

[0028]

[0029] In the formula, CTR, CR and USR represent click-through rate, conversion rate and user satisfaction rate respectively. z , CR z and USR z They represent the benchmark values ​​of click-through rate, conversion rate, and user satisfaction rate respectively; a, b, and c represent adjustment coefficients, and their value ranges are all 0 to 1.

[0030] Furthermore, when AI user behavior prediction promotion is adopted, the second evaluation index is collected, including click-through rate, conversion rate and user retention rate, and an evaluation formula is established based on the second evaluation index to calculate the effect evaluation index Tg;

[0031] The evaluation formula established based on the second evaluation indicator is as follows:

[0032]

[0033] In the formula, URR represents user retention rate, URR z It represents the base value of user retention rate. d, e and f represent adjustment coefficients, and their values ​​range from 0 to 1.

[0034] Furthermore, the process of adjusting the established promotion frequency in real time according to the generated effect evaluation index is as follows:

[0035] The effect evaluation index is compared with the preset standard threshold. If the effect evaluation index exceeds the standard threshold, the established promotion frequency is increased according to the set value; if the effect evaluation index does not exceed the standard threshold, the established promotion frequency is reduced according to the set value and the optimization strategy is executed.

[0036] Furthermore, the optimization strategy implemented is: if it is aimed at interactive content creation promotion, then optimize the interactive content; if it is aimed at AI user behavior prediction promotion, then optimize the prediction model.

[0037] A new media promotion optimization method based on the Internet includes the following steps:

[0038] S1. Collect and analyze multi-source Internet data to build user portraits;

[0039] S2. Based on user portraits, use NLP and GAN technology to generate corresponding promotional content;

[0040] S3. Based on user portraits and historical promotion data, a deep learning model is introduced to predict the user groups interested in the promotion content;

[0041] S4. Based on the key indicators reflecting user activity, a rule engine is built to identify the corresponding users as active users or inactive users, and the promotion is performed by the method set in the solution database according to the user type;

[0042] S5. Under the promotion method corresponding to the user type, the effect of each promotion is evaluated, and the established promotion frequency is adjusted in real time based on the generated effect evaluation index.

[0043] (III) Beneficial effects

[0044] The present invention provides an Internet-based new media promotion optimization system and method, which has the following beneficial effects:

[0045] (1) The solution introduces a rule engine to define active users and inactive users, and sets adjustable definition rules. This enables the system to be flexibly adjusted according to different business needs and changes in user behavior, ensuring the effectiveness and adaptability of the promotion strategy. At the same time, by regularly collecting user feedback and business data, the rule engine and promotion strategy are evaluated and improved, achieving continuous optimization and iteration of the system.

[0046] (2) By establishing clear evaluation indicators and evaluation formulas, the scheme can comprehensively and accurately measure the effectiveness of each promotion activity. For different promotion methods, the most representative indicators are selected for evaluation to ensure the accuracy of the evaluation results. Based on the comparison between the effect evaluation index and the preset standard threshold, the system can automatically and real-time adjust the promotion frequency, which not only avoids the waste of resources, but also ensures the continuous optimization of promotion activities and improves the promotion efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a modular schematic diagram of a new media promotion optimization system based on the Internet in the present invention. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] Embodiment 1:

[0050] See also Figure 1 This embodiment provides an Internet-based new media promotion optimization system, which implements differentiated promotion strategies for different types of users (active users and inactive users) and dynamically adjusts the promotion frequency according to the real-time evaluation value of the promotion effect to maximize the promotion effect; the system includes functional modules that run in sequence, namely, a multi-dimensional data analysis module, an intelligent content creation module, a user positioning promotion module, a user-defined secondary promotion module, and an effect evaluation and adjustment module;

[0051] Multi-dimensional data analysis module, which collects and analyzes multi-source Internet data to build user portraits, including but not limited to user interest preferences, browsing habits, and interaction history;

[0052] The multi-dimensional data analysis module is an advanced system that integrates big data processing capabilities. Its core function is to collect and analyze data from multiple sources to build detailed user portraits. This module can not only help companies understand their user groups more deeply, but also provide solid data support for precision marketing strategies.

[0053] The following is a detailed analysis of the module's functions:

[0054] Sources of Internet multi-source data:

[0055] 1. Social media: User public information: such as personal profiles, published posts, shared links, etc., which can reflect users' interests, opinions and lifestyles; social interactions: interactive behaviors such as likes, comments, and reposts, which reveal users' preferences and participation in different content;

[0056] 2. Search engine: Search keywords: The search terms entered by users directly reflect their current needs and interests; Search history: Long-term search records can reveal the topics and trends that users continue to pay attention to;

[0057] 3. User behavior logs: Browsing history: The browsing path of users on websites or applications shows their interest in different content or products; Click behavior: Interactive behaviors such as clicking links and buttons can analyze users' reactions and preferences to specific content; Purchase records: For e-commerce platforms, users' purchase history is an important basis for understanding their consumption habits and preferences;

[0058] Data analysis and user portrait construction:

[0059] Data integration: Clean, deduplicate and integrate Internet multi-source data from different sources to form a unified data set, laying the foundation for subsequent analysis;

[0060] Feature extraction: Extract key features from the integrated data, such as user interest preferences, browsing habits, interaction patterns, etc. These features can be explicit (such as interest information directly provided by users) or implicit (such as interests inferred through behavioral analysis);

[0061] Application of machine learning algorithms: Use machine learning algorithms such as clustering, classification, and regression to conduct in-depth analysis of the extracted features; identify subtle differences between user groups, such as the behavior patterns of user groups of different ages, regions, or interests;

[0062] User portrait construction: Based on the above analysis, a multi-dimensional user portrait is constructed, including the user's interest preferences, browsing habits, and interaction history. The user portrait can be static (such as the user's basic information) or dynamic (such as the changes in the user's interest preferences over time).

[0063] In summary, advanced machine learning algorithms are used to conduct in-depth mining of massive data to discover potential user needs and interests. Through sophisticated data analysis, subtle differences between different user groups can be identified, providing the possibility for more accurate promotion strategies. User portraits are dynamically constructed and can be updated in real time as user behavior changes, ensuring the timeliness and accuracy of promotion strategies. Therefore, the multi-dimensional data analysis module integrates multi-source data and applies machine learning algorithms for in-depth mining and analysis to construct detailed user portraits, providing strong data support for the company's precise promotion and personalized services.

[0064] Intelligent content creation module, based on user portraits, uses NLP and GAN technology to generate corresponding promotional content;

[0065] This module receives the output results of the multi-dimensional data analysis module, which includes key information such as user portraits, user interest preferences, browsing habits, and interaction history. Based on these analysis results, the module can accurately locate the needs and interests of the target user group;

[0066] Using NLP technology, it is possible to automatically generate copywriting that is consistent with the brand tone and highly matched with the interests of target users. These copies can be in various forms such as advertising slogans, product descriptions, and press releases. Through GAN technology, it is possible to generate creative and attractive image and video content, which not only complements the copywriting, but also further attracts the user's attention through visual elements.

[0067] In addition, in addition to generating new content, it is also possible to initially optimize existing content;

[0068] For example, we can polish the copy, edit and process images and videos with special effects to improve the overall promotion effect. Through deep learning and algorithm models, we can continuously learn and adjust the strategy of generating content to ensure that the content is always highly matched with the interests of the target user group.

[0069] It should be noted that:

[0070] Natural Language Processing (NLP) technology: NLP technology enables the module to understand the structure and meaning of human language, so that it can automatically generate copywriting that conforms to grammatical and semantic rules. By analyzing user portraits and interest preferences, NLP technology can also enable the module to generate more personalized and targeted copywriting content;

[0071] Generative Adversarial Network (GAN) technology: GAN technology can generate creative and diverse image and video content through the adversarial training process of the generator and the discriminator. In the intelligent content creation module, GAN technology is used to generate visual content that is highly matched with the interests of the target user group and in line with the brand tone;

[0072] Creativity and Personalization: The module can not only generate standard content that meets the brand tone, but also realize the creativity and personalization of content through the innovative application of algorithm models; for example, it can generate creative copywriting expressions through NLP technology, or generate unique image and video styles through GAN technology to meet the personalized needs of different users;

[0073] Maintaining brand consistency: When generating or optimizing promotional content, the module can fully consider the consistency requirements of brand tone. Through the preset brand tone parameters and the constraints of the algorithm model, the module can ensure that the generated content not only meets the interests of the target user group but also maintains the brand characteristics.

[0074] To sum up, the intelligent content creation module is an automated promotion content generation system based on multi-dimensional data analysis results and advanced technologies. It can achieve the creativity and personalized customization of content through the innovative application of NLP and GAN technologies, while maintaining the consistency of brand tone, providing strong support for the company's precise promotion and personalized services.

[0075] User positioning promotion module, based on user portraits and historical promotion data, introduces a deep learning model to predict the user groups interested in the promotion content;

[0076] This module is a system component based on advanced technologies and algorithms, which aims to accurately identify the user groups most likely to be interested in specific promotional content by analyzing rich user data; this process not only improves the efficiency of marketing promotion, but also ensures that resources can be allocated more targetedly, thereby maximizing the return on investment (ROI); The following is a detailed description of this module in combining deep learning models, especially deep neural networks (DNN) and graph neural networks (GNN), to achieve precise user positioning:

[0077] User portrait:

[0078] Collect multi-dimensional data including basic user information (such as age, gender, geographic location), interest preferences, purchase history, browsing behavior, etc. to build a comprehensive user portrait;

[0079] Historical promotion data:

[0080] Organize the data of past promotional activities, including promotional content, audience characteristics, response rate, conversion rate, etc., as an important reference for model training;

[0081] The process of introducing a deep learning model to predict the user group interested in the promoted content is as follows:

[0082] Feature Engineering:

[0083] Extract key features from user profiles and historical promotion data. These features should reflect user preferences, behavior patterns, and potential associations with promotional content. For example, the user profiles and historical promotion data described above can be used to select features using statistical methods or machine learning techniques (such as PCA and LDA), remove redundant information, and retain the most valuable features for prediction, so as to reduce model complexity.

[0084] Introducing deep learning models:

[0085] Construct a multi-layer neural network. The input layer receives preprocessed feature data, the hidden layer learns deep feature interactions through nonlinear transformation, and the output layer predicts the probability of user interest in the promoted content. Use large-scale data sets for model training, and use back-propagation algorithms and optimizers (such as Adam and SGD) to adjust network weights and minimize prediction errors. DNN can automatically learn complex feature representations, capture nonlinear relationships in data, and improve prediction accuracy.

[0086] Model evaluation and tuning:

[0087] Use methods such as K-fold cross validation to evaluate model performance, ensure the stability and generalization ability of the model on different data sets, pay attention to indicators such as accuracy, recall, F1 score, AUC-ROC curve, comprehensively evaluate the model effect, and adjust the model structure (such as the number of layers, number of neurons), learning rate, regularization parameters, etc. according to the evaluation results to optimize the model performance;

[0088] Precise push and effect monitoring:

[0089] Based on the model prediction results, screen out the user groups most likely to be interested in the promoted content; based on user preferences and prediction results, customize personalized promotion strategies to improve user experience; monitor the effectiveness of promotion activities in real time and collect feedback data for subsequent model iteration and optimization.

[0090] The user-defined secondary promotion module builds a rule engine based on key indicators that reflect user activity, identifies the corresponding users as active or inactive users, and uses the method set in the solution database to promote according to the user type;

[0091] Among them, key indicators include login frequency, usage time, and interactive behavior;

[0092] Login frequency: number of logins per day / week / month;

[0093] Duration of use: duration of use each time / day / week / month;

[0094] Interaction behavior: the number of purchase behaviors per month;

[0095] The process of building a rule engine to identify the corresponding users as active users or inactive users is as follows:

[0096] Set the AND logic mechanism in the rule engine and mark users as active users if they meet the following conditions:

[0097] Daily active users: Users must log in at least once a day;

[0098] Weekly active users: Users must log in at least 3 times a week, and use the app for more than 20 minutes each time;

[0099] Monthly active users: Users must make at least 5 purchases per month;

[0100] In this case, only if a user meets all the listed conditions within a given time period, they will be marked as active; otherwise, they are marked as inactive;

[0101] The rule engine can use simple logical judgments or complex algorithm models;

[0102] Here is a basic rules engine example:

[0103]

[0104]

[0105] In addition, user feedback and business data are collected regularly to evaluate and improve the effectiveness of the rule engine. The optimal definition rules are found through methods such as A / B testing. Through the above steps, a flexible rule engine can be built to define active and inactive users, and the rules can be continuously optimized and adjusted according to business needs. The details are not described here.

[0106] When the promotion is performed by calling the method set in the solution database according to the user type, the solution database includes interactive content creation promotion and AI user behavior prediction promotion;

[0107] When the user type is: active user, the promotion method used is interactive content creation promotion;

[0108] When the user type is: inactive user, the promotion method used is AI user behavior prediction promotion;

[0109] It should be noted that interactive content creation and promotion:

[0110] Solution description: Incorporate interactive elements (such as voting, Q&A, mini-games, etc.) into content creation to enable the audience to directly participate in the content, making it more interesting and attractive. For example, you can create a brand-related online mini-game so that users can understand and experience the brand by participating in the game. Targeted scenarios: Suitable for brands that want to increase user engagement and interactivity, especially for active user groups.

[0111] AI user behavior prediction promotion:

[0112] Solution description: Using advanced natural language processing and machine learning technologies, automatically generate content that is consistent with the brand tone and preferences of the target audience. This can not only improve the efficiency of content creation, but also ensure the personalization and differentiation of content to meet the diverse needs of users; Implementation points: It is necessary to collect and analyze a large amount of user data, train AI models to understand user preferences and content trends, and at the same time, maintain manual review and optimization of content to ensure the authenticity and value of the content.

[0113] By adopting the above technical solution, the following significant technical effects are achieved:

[0114] Accurate user positioning and efficient promotion:

[0115] By using deep learning models such as deep neural networks (DNN) and graph neural networks (GNN), combined with rich user portraits and historical promotion data, we can accurately locate users. By deeply mining user preferences, behavior patterns and potential associations with promotional content, we can accurately identify the user groups most likely to be interested in specific promotional content, thereby improving the efficiency and accuracy of marketing promotion.

[0116] Resource optimization and ROI maximization:

[0117] Through precise user positioning, resources can be allocated more specifically, avoiding the blindness of casting a wide net. This ensures that every investment can generate the greatest return, thus maximizing the return on investment (ROI);

[0118] Personalized promotion and improved user experience:

[0119] Different promotion strategies are adopted for different user types (active users and inactive users). For active users, interactive content creation promotion is adopted. Through interactive elements such as voting, Q&A, and mini-games, the content becomes more interesting and attractive, further improving user participation and interactivity. For inactive users, AI user behavior prediction promotion is adopted. Machine learning technology is used to generate personalized content that meets user preferences, meeting the diverse needs of users and improving the efficiency of content creation.

[0120] Flexibility and Optimizability:

[0121] The system introduces a rule engine to define active users and inactive users, and sets adjustable definition rules, which enables the system to flexibly adjust according to different business needs and changes in user behavior, ensuring the effectiveness and adaptability of the promotion strategy; at the same time, by regularly collecting user feedback and business data, evaluating and improving the rule engine and promotion strategy, the system is continuously optimized and iterated;

[0122] Data-driven decision-making and intelligent operations:

[0123] The entire system is built based on big data and machine learning technologies, realizing data-driven decision-making and intelligent operations. Through in-depth analysis and mining of user data, it can gain insights into user behaviors and preferences, providing strong support for formulating more effective promotion strategies. At the same time, the automation and intelligent features of the system also reduce labor costs and improve operational efficiency.

[0124] To sum up, by adopting the above technical solutions, we can not only achieve accurate user positioning and efficient promotion, but also optimize resource allocation, improve user experience, enhance system flexibility and optimizability, and promote data-driven decision-making and intelligent operations. These technical effects together bring significant benefits and competitive advantages to market promotion.

[0125] The effect evaluation and adjustment module evaluates the effect of each promotion under the promotion method corresponding to the user type, and adjusts the established promotion frequency in real time based on the generated effect evaluation index;

[0126] The evaluation process of each promotion effect is as follows:

[0127] When interactive content creation and promotion is adopted, the first evaluation index is collected, including click-through rate, conversion rate and user satisfaction rate, and an evaluation formula is established based on the first evaluation index to calculate the effect evaluation index;

[0128] The evaluation formula established based on the first evaluation indicator is as follows:

[0129]

[0130] In the formula, CTR, CR and USR represent click-through rate, conversion rate and user satisfaction rate respectively. z , CR z and USR z They represent the benchmark values ​​of click-through rate, conversion rate, and user satisfaction rate, i.e., the average values ​​of historical industry data. a, b, and c represent adjustment coefficients, all ranging from 0 to 1, which are used to adjust the influence of each indicator on the overall evaluation index. These coefficients can be determined through experiments or experience to reflect the importance of different indicators in the evaluation.

[0131] Formula design logic:

[0132] The numerator (CTR×CR×USR) reflects the combined effect of the three indicators, namely the product of clicks, conversions and user satisfaction rate, and represents a comprehensive measure of success. The denominator compares the actual value with the benchmark value and introduces an adjustment coefficient to reflect the improvement or decline of each indicator relative to the benchmark. If an indicator performs better than the benchmark, its contribution to the denominator will decrease, thereby improving the overall evaluation index. On the contrary, if the performance is poor, it will lower the evaluation index.

[0133] When AI user behavior prediction promotion is adopted, the second evaluation index is collected, including click-through rate, conversion rate and user retention rate, and an evaluation formula is established based on the second evaluation index to calculate the effect evaluation index;

[0134] The evaluation formula established based on the second evaluation indicator is as follows:

[0135]

[0136] In the formula, URR represents user retention rate, URR z It represents the benchmark value of user retention rate, that is, the average value of industry historical data. d, e, and f represent adjustment coefficients, all ranging from 0 to 1, reflecting the importance of each indicator.

[0137] Formula design logic:

[0138] They respectively represent the degree of improvement of click-through rate, conversion rate and user retention rate relative to the benchmark values; the last factor takes into account the coordination between the three indicators. By calculating the difference between the click-through rate, conversion rate and user retention rate, and dividing them by the sum of their benchmark values, we can get a value that reflects the degree of incoordination between the indicators. The smaller this value is, the more coordinated the indicators are, thereby improving the overall evaluation index.

[0139] Further explanation is as follows:

[0140] Common evaluation indicators: click-through rate, conversion rate:

[0141] Whether it is interactive content creation promotion or AI user behavior prediction promotion, click-through rate is a basic indicator to measure whether users are interested in the promotion content; it reflects the attractiveness of the promotion content and the initial reaction of users; the conversion rate measures whether users take further expected actions (such as purchase, registration, download, etc.) after clicking, and this indicator is crucial for evaluating the promotion effect because it is directly related to the commercial goals of the promotion;

[0142] Unique evaluation indicators:

[0143] User satisfaction rate (only for active users)

[0144] Reason: Interactive content creation promotion often emphasizes direct interaction with users and personalized experience; therefore, user satisfaction rate becomes an important indicator for evaluating the effectiveness of such promotion; satisfaction reflects the user's preference for interactive content, the quality of experience, and whether it meets their needs; for active users, they are more willing to participate and give feedback, so user satisfaction rate can more accurately reflect the promotion effect;

[0145] User retention rate (only for inactive users)

[0146] Reason: For inactive users, AI user behavior prediction promotion mainly predicts their future needs by analyzing their historical behavior and promotes accordingly. Since these users are not active enough, it is particularly important to measure whether they become more active or stay on the platform due to promotion. User retention rate reflects the effect of promotion on improving user stickiness and activity, and is a key indicator for evaluating the success of such promotion.

[0147] In summary, click-through rate and conversion rate, as basic indicators for measuring promotion effectiveness, are common to both promotion methods; while user satisfaction rate and user retention rate are evaluated based on the unique characteristics of interactive content creation promotion and AI user behavior prediction promotion, respectively, and therefore are only used as evaluation indicators in the corresponding promotion methods;

[0148] The process of adjusting the established promotion frequency in real time based on the generated effect evaluation index is as follows:

[0149] The effect evaluation index is compared with the preset standard threshold. If the effect evaluation index exceeds the standard threshold, it indicates that the promotion effect is good, and the established promotion frequency is increased according to the set value (for example, if the original established promotion frequency is twice a day and the set value is 1, the increased promotion frequency is three times a day) until the effect evaluation index does not exceed the standard threshold; if the effect evaluation index does not exceed the standard threshold, it indicates that the promotion effect is poor, and the established promotion frequency is reduced according to the set value, and the optimization strategy is executed; among them, the preset standard threshold is set according to the actual situation, and no further details are given here;

[0150] The optimization strategies implemented are:

[0151] If you are promoting interactive content, optimize the interactive content;

[0152] For example: Enhance interactivity: introduce more interactive elements, such as voting, Q&A, and mini-games, to allow users to participate more actively; use dynamic content or personalized recommendations to customize the interactive experience based on user behavior or interests; improve content quality: ensure that the content is accurate, interesting, and educational to attract users' attention; use high-quality visual and audio elements to enhance the overall user experience;

[0153] If the promotion is based on AI user behavior prediction, optimize the prediction model;

[0154] For example: Algorithm improvement: try different machine learning algorithms, such as random forest, gradient boosting tree, neural network, etc., to find the model that best suits the current data; adjust the algorithm parameters, such as learning rate, number of iterations, regularization terms, etc., to optimize model performance.

[0155] By adopting the above technical solution, the following technical effects are achieved:

[0156] Accurate effect evaluation: This solution can comprehensively and accurately measure the effect of each promotion activity by setting up clear evaluation indicators and evaluation formulas. It selects the most representative indicators for evaluation of different promotion methods to ensure the accuracy of the evaluation results.

[0157] Dynamically adjust promotion frequency: Based on the comparison between the effect evaluation index and the preset standard threshold, the system can automatically and in real time adjust the promotion frequency, which not only avoids the waste of resources, but also ensures the continuous optimization of promotion activities and improves the promotion efficiency;

[0158] Personalized promotion strategy: For interactive content creation promotion and AI user behavior prediction promotion, unique evaluation indicators are set, such as user satisfaction rate and user retention rate, which helps to formulate more personalized promotion strategies to better meet the needs of different user groups;

[0159] Continuous optimization and iteration: Through the feedback of effect evaluation index, the system can promptly identify problems in promotion activities and trigger corresponding optimization strategies. Whether it is optimizing interactive content or predictive models, it helps to promote the continuous optimization and iteration of promotion activities and improve the overall promotion effect.

[0160] Improve user stickiness and conversion rate: By accurately positioning users, formulating personalized promotion strategies, and continuously optimizing promotion activities, we can effectively improve user stickiness and conversion rate, bringing significant commercial value to the company's marketing activities;

[0161] In summary, this technical solution achieves optimization and iteration of promotion activities through accurate effect evaluation and dynamic adjustment of promotion frequency, improves promotion efficiency and user conversion rate, and brings significant market competitive advantages to enterprises.

[0162] Embodiment 2:

[0163] Based on Example 1, this embodiment also provides an Internet-based new media promotion optimization method, including the following specific steps:

[0164] S1. Collect and analyze multi-source Internet data to build user portraits;

[0165] S2. Based on user portraits, use NLP and GAN technology to generate corresponding promotional content;

[0166] S3. Based on user portraits and historical promotion data, a deep learning model is introduced to predict the user groups interested in the promotion content;

[0167] S4. Based on the key indicators reflecting user activity, a rule engine is built to identify the corresponding users as active users or inactive users, and the promotion is performed by the method set in the solution database according to the user type;

[0168] S5. Under the promotion method corresponding to the user type, the effect of each promotion is evaluated, and the established promotion frequency is adjusted in real time based on the generated effect evaluation index.

[0169] In the application, the several formulas involved are all calculated by taking their numerical values ​​after removing the dimensions, and the formulas are established by collecting a large amount of data and performing software simulation to obtain a formula for the most recent actual situation. Some coefficients or weights in the formulas are set by technical personnel in this field according to actual conditions, so they will not be elaborated here.

[0170] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product. A person of ordinary skill in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution.

[0171] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0172] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A new media promotion optimization system based on the Internet, characterized in that: The system includes: Multi-dimensional data analysis module, which collects and analyzes multi-source data on the Internet to build user portraits; Intelligent content creation module, based on user portraits, uses NLP and GAN technology to generate corresponding promotional content; User positioning promotion module, based on user portraits and historical promotion data, introduces a deep learning model to predict the user groups interested in the promotion content; The user-defined secondary promotion module builds a rule engine based on key indicators reflecting user activity, identifies the corresponding user as an active user or an inactive user, and uses the method set in the solution database to promote according to the user type; key indicators include login frequency, usage time, and interactive behavior; where interactive behavior refers to the number of purchases per month; when promoting according to the method set in the solution database based on user type, the solution database includes interactive content creation promotion and AI user behavior prediction promotion; When the user type is: active user, the promotion method used is interactive content creation promotion; When the user type is: inactive user, the promotion method used is AI user behavior prediction promotion; The effect evaluation and adjustment module evaluates the effect of each promotion under the promotion method corresponding to the user type, and adjusts the established promotion frequency in real time based on the generated effect evaluation index.

2. The Internet-based new media promotion optimization system according to claim 1, characterized in that: The user profile constructed includes but is not limited to user interest preferences, browsing habits, and interaction history.

3. The new media promotion optimization system based on the Internet according to claim 2 is characterized in that: The process of building a rule engine to identify the corresponding users as active users or inactive users is as follows: Set the AND logic mechanism in the rule engine and set the conditions as follows: Daily active users: users log in at least once a day; Weekly active users: users log in at least 3 times a week, and each session lasts more than 20 minutes; Monthly active users: users make at least 5 purchases per month; When the corresponding user meets all conditions within a given time period, he is marked as an active user; otherwise, he is marked as an inactive user.

4. The Internet-based new media promotion optimization system according to claim 1, characterized in that: The evaluation process for each promotion effect is as follows: When interactive content creation and promotion is adopted, the first evaluation index is collected, including click-through rate, conversion rate and user satisfaction rate, and an evaluation formula is established based on the first evaluation index to calculate the effect evaluation index Tg; The evaluation formula established based on the first evaluation indicator is as follows: In the formula, CTR, CR and USR represent click-through rate, conversion rate and user satisfaction rate respectively. z , CR z and USR z They represent the benchmark values ​​of click-through rate, conversion rate, and user satisfaction rate respectively; a, b, and c represent adjustment coefficients, and their value ranges are all 0 to 1.

5. The Internet-based new media promotion optimization system according to claim 4 is characterized in that: When AI user behavior prediction promotion is adopted, the second evaluation index is collected, including click-through rate, conversion rate and user retention rate, and an evaluation formula is established based on the second evaluation index to calculate the effect evaluation index Tg; The evaluation formula established based on the second evaluation indicator is as follows: In the formula, URR represents user retention rate, URR z It represents the base value of user retention rate. d, e and f represent adjustment coefficients, and their value ranges are all 0 to 1.

6. The Internet-based new media promotion optimization system according to claim 5, characterized in that: The process of adjusting the established promotion frequency in real time based on the generated effect evaluation index is as follows: The effect evaluation index is compared with the preset standard threshold. If the effect evaluation index exceeds the standard threshold, the established promotion frequency is increased according to the set value; if the effect evaluation index does not exceed the standard threshold, the established promotion frequency is reduced according to the set value and the optimization strategy is executed.

7. The Internet-based new media promotion optimization system according to claim 6, characterized in that: The optimization strategy implemented is: if it is for interactive content creation promotion, then optimize the interactive content; if it is for AI user behavior prediction promotion, then optimize the prediction model.

8. A new media promotion optimization method based on the Internet, using any system described in claims 1 to 7, characterized in that: The steps include: S1. Collect and analyze multi-source Internet data to build user portraits; S2. Based on user portraits, use NLP and GAN technology to generate corresponding promotional content; S3. Based on user portraits and historical promotion data, a deep learning model is introduced to predict the user groups interested in the promotion content; S4. Based on the key indicators reflecting user activity, a rule engine is built to identify the corresponding users as active users or inactive users, and the promotion is performed by the method set in the solution database according to the user type; Key indicators include login frequency, usage time, and interactive behavior; interactive behavior refers to the number of purchases per month; when promoting based on user type using the method set in the solution database, the solution database includes interactive content creation promotion and AI user behavior prediction promotion; When the user type is: active user, the promotion method used is interactive content creation promotion; When the user type is: inactive user, the promotion method used is AI user behavior prediction promotion; S5. Under the promotion method corresponding to the user type, the effect of each promotion is evaluated, and the established promotion frequency is adjusted in real time based on the generated effect evaluation index.

Images